C h a p t e r 7
Results are presented of "shear" fracture tests on concrete. Two different but unique test methods were used. In the first approach, double-edge-notched quadratic plates are sub jected to combined tension and shear. During loading the boundaries of the plates are kept parallel. Depending on the exact boundary conditions and the load-path, a set of curved or straight overlapping cracks develops from the two notches The behaviour seems governed by mode I cracking only. Beam tests are carried out in a sophisticated apparatus where the boundary conditions can be varied. Tests can either be carried out between fixed or rotating supports. It is found that the actual boundary condition has a significant effect on the crack mode. Under fixed supports, the overlapping curved crack mode is found, and final failure occurs through splitting. This situation resembles the results obtained in the plate experi ments. When on the other hand the supports in the beam tests are allowed to rotate, a single mode I crack dominates the behaviour of the single-edge-notched and double-edgenotched four point shear beams. The results are confirmed by simulations with a simple lattice model, which are presented in another contribution to this conference . Shear failure, defined as an array of inclined microscopic tensile cracks seems only be possible when secondary cracking can be prevented (1) through confinement on the parts of the specimen outside the expected shear zone, (2) under dynamic loading, or (3) in fibre reinforced cement composites where secondary cracking is prevented by the fibres.